brian@natinst.com (Brian H. Powell) (08/22/90)
We've been running Cap 5.0 on a Sun3 for a few years without too much
problem. Now, I'm trying to get Cap+ to run on a Sun 4/330 running SunOS 4.1.
Things seem to be going pretty well, but atlook sometimes dumps core.
I've never been to gung ho about debugging CAP, so I'm hoping this will
ring a bell with somebody, or somebody can help me figure it out.
Here's the stack trace:
#0 0x5924 in nbp_match (pr=(struct nbpProto *) 0x2015efd, id=0 '\000') (abnbp.c line 574)
#1 0x6de8 in q_mapf (...) (...)
#2 0x57bc in LkUpReply (nbp=(NBP *) 0x15b98, len=74) (abnbp.c line 531)
#3 0x5570 in nbp_listener (skt=0 '\000', type=91 '[', nbp=(NBP *) 0x15b98, len=74, addr=(AddrBlock *) 0xf7ffcccc) (abnbp.c line 436)
#4 0x3dd4 in ddp_protocol (iov=(struct iovec *) 0x127f0, iovlen=1, plen=74) (abddp.c line 410)
#5 0x92f4 in kip_get (...) (...)
#6 0x7c08 in fdlistenread (...) (...)
#7 0x8648 in abSelect (...) (...)
#8 0x8358 in abSleep (...) (...)
#9 0x51a8 in nbpFcn (abr=(struct nbpProto *) 0xf7fffd58, fcn=21, async=0) (abnbp.c line 313)
#10 0x4d9c in NBPLookup (abr=(struct nbpProto *) 0xf7fffd58, async=0) (abnbp.c line 139)
#11 0x2c20 in dolookup ()
#12 0x2abc in main ()
As those of you familiar with the source can see, LkUpReply calls
q_mapf() to traverse the NBP queue. Each element of the queue is passed to
nbp_match to see if the ID matches. It appears that one of the addresses in
the queue has one of the high bits set, and this causes a seg violation.
As I mentioned above, this doesn't always happen. I'd guess about half
or a third of the time. Thanks for any help.
Brian H. Powell, M/S 56-14 National Instruments Corp.
brian@natinst.com 6504 Bridge Point Parkway
uunet!cs.utexas.edu!natinst!brian Austin, Texas 78730-5039
AppleLink:NATINST (512) 338-9119